Fix standardize trim_table argument

Small fix on the standardize method. Currently, standardize with trim_table=True wont work without the mapper argument set. This merge request fixes this issue.

You can test this bug with the following code (it will throw an error without this branch):

import pandas as pd
from ketos.data_handling import selection_table as sl

annot = pd.DataFrame([{"filename":"sample_audio.wav", "start":2.0, "end":3.0, "label":"A", "duration":1, "offset":50},
                     {"filename":"sample_audio.wav", "start":5.0, "end":6.0, "label":"A",  "duration":1, "offset":50},
                     {"filename":"sample_audio.wav", "start":21.0, "end":22.0, "label":"A",  "duration":1, "offset":50},
                     {"filename":"sample_audio.wav", "start":25.0, "end":27.0, "label":"A",  "duration":1, "offset":50}])


annot_std = sl.standardize(table=annot, labels=["A"], trim_table=True, start_labels_at_1=True)

annot_std

docs/source/versions/changelog.rst

 Change log
 ==========
 
+**Version 2.6.2** (Month Day, Year)
+
+ * Fixed bug in :meth:`standardize <ketos.data_handling.selection_table.standardize>` where setting `trim_table=True` and `mapper=None` would give an error.
+
 **Version 2.6.1** (June 3, 2022)
 
  * In :meth:`AudioWriter:close <ketos.data_handling.database_interface.AudioWriter.close>` remove any existing column indices before attempting to reindex them. (Only relevant if appending to an existing database.)

ketos/data_handling/selection_table.py

@@ -267,8 +267,11 @@ def standardize(table=None, path=None, sep=',', mapper=None, labels=None,
 
     # keep only relevant columns
     if trim_table:
-        df = trim(df, list(mapper.keys()))
-
+        if mapper is not None:
+            df = trim(df, list(mapper.keys()))
+        else:
+            df = trim(df)
+            
     # check that dataframe has minimum required columns
     mis = missing_columns(df)
     assert len(mis) == 0, 'Column(s) {0} missing from input table'.format(mis)